Here's the updated draft.
Hi Inbox Detox working group,
A quick reminder that our working group meeting is Wednesday, July 8, at 11:00 AM Eastern (EDT). For those in other zones:
- 9:00 AM MDT (US Mountain)
- 8:00 AM PDT (US Pacific)
- 8:30 PM IST (India)
We'll meet on Webex: https://ieee.webex.com/meet/awalter
On the agenda: Aubrey and Mary are planning to provide an update on the Jira stories they're writing for the preferences center and the digest, and they're looking for review and input from the group.
I'll also give a live demo of a working prototype I've built of the eNotice Digest. It's modeled on the mockup we put together during Phase 2, but instead of a static page it generates each digest dynamically from a personalized unit selection. A few details on it:
Key Features:
- Personalized Selection: Starting from a member's IEEE Section and Societies (if any), the app walks the public OU List API to find their related units — parent and child organizations including Regions, Chapters, and Affinity Groups, among others — and surfaces the eNotices sent to the ones they follow within a chosen date window.
- AI Summaries & Tags: A Claude model writes a neutral 2–3 sentence summary of each eNotice and generates topical hashtags drawn from the content, the IEEE Taxonomy, the associated event, and the location. Hover any tag to see exactly how it was derived.
- Grounded Event Details: For eNotices tied to an event, it pulls structured data from the vTools Events API (title, tags, location, image) rather than scraping pages, keeping the summaries and tags accurate.
- Two Views: A tiled Digest for at-a-glance reading, plus a sortable, filterable Archive table (with per-unit "received" counts) for digging into the full set.
Please note this is a proof of concept focused on the member-facing experience. It runs on a snapshot of exported eNotice data (not a live feed), and the summaries, images, and tags are generated on demand when you open a digest — so give it a few seconds per view. A production version would prepare that content once when each eNotice is published, and would add live eNotice data and IEEE SSO; the tag rules and taxonomy are also still being refined.
You can try it here: https://enotice-digest-prototype.streamlit.app/ (Access password: AdvancingTechnologyForHumanity). Heads up: it's currently running on my personal Anthropic and OpenAI API accounts seeded for testing, so feel free to click around, but please hold off on sharing the link or password outside this group for now.
If you'd like a look behind the scenes, the code is in vTools Labs: https://opensource.ieee.org/vtoolslabs/enotice-digest-prototype
Talk to you soon!
Thanks,
Chris